A field study of phosphorus mobilisation from commercial fertilisers.
Agricultural catchments contribute to excessive phosphorus (P) concentrations, and associated algal problems, in the waterways of south-eastern Australia (Gutteridge et al. 1992; Environment Protection Authority 1995; Webster et al. 2001). In this region, P fertilisers are used extensively to enhance the productivity of ryegrass (Lolium spp.) and clover (Trifolium spp.) based pastures that are grazed year-round (Connor and Smith 1987).
Fertilisers applied to agricultural land contribute to the export of P from pasture (Sharpley et al. 1978; Olness et al. 1980; Preedy et al. 2001; Withers et al. 2001). These studies suggest that the form of the fertiliser, including its solubility, is important. For example, more total P (TP) was exported from plots receiving dicalcium phosphate (CaHP[O.sub.4]) than from plots receiving the more soluble monocalcium phosphate [MCP, Ca[([H.sub.2]P[O.sub.4]).sub.2]] (Sharpley et al. 1978). However, the reverse was true for dissolved P, reflecting the solubility of MCR
The timing of P additions also affects P exports. The first overland flow (runoff) event following P additions generally contains the highest P concentrations (Edwards and Daniel 1994; Austin 1998). More importantly, as the time between P application and the first overland flow event increases, P concentrations decrease (Sharpley 1997; Nash et al. 2000).
Pasture systems in south-eastern Australia are sustained by either natural rainfall (rainfed) or rainfall supplemented with irrigation from sources such as groundwater or water impoundments. Irrigation systems vary from spray irrigation that has many of the attributes of rainfall, to the more common border (also called border-check, flood or surface) irrigation. In border irrigation, water is applied to the top of a bay (commonly c. 350 by 40 m) in excess of the soil infiltration rate and moves down the bay as infiltration excess overland flow (Nash et al. 2002). Water in border irrigation systems is confined by check banks (i.e. raised earthen ridges) down the sides of the bay. Excess water, c. 20% of that applied (Nexhip et al. 1997), is removed from the foot of the bay into drainage channels. To assist the even distribution of water within the bay, most border irrigation systems are laser graded to a slope of between 1:400 and 1:1000, depending on the region. Water is usually applied 10-20 times from late spring to early autumn, with annual applications of between 5 and 10 ML/ha.
In Australia the P fertilisers most often applied to pastures are MCR predominantly in the form of single superphosphate [SSP, Ca[([H.sub.2]P[O.sub.4]).sub.2]] and diammonium phosphate [DAR (N[H.sub.4])2HP[O.sub.4]]. Previous studies have shown that SSP is less soluble than DAP (Nash et al. 2003a) and this may affect P concentrations in overland flow.
The aim of this study was to compare P exports in irrigation and rainfall-induced overland flow following the surface application of DAP and SSP to field soils. This information could then be used to develop guidelines that minimise P exports where fertilisers are applied when overland flow might reasonably be expected in the following 4-10 days.
Materials and methods
Phosphorus concentrations in overland flow following the broadcasting (or surface application) of 2 commercial P fertilisers, DAP (Incitec Pivot Ltd, Geelong, Vic., Australia) and SSP (SuPerfect[TM], Incitec Pivot Ltd) were investigated in 2 experiments. To balance the nutrients added with the fertilisers [i.e. nitrogen (N) in DAP and gypsum, CaS[O.sub.4] in SSP], fertiliser blends that supplied 50 kg P/ha and 80 kg N/ha were applied in both studies. Urea and ammonium sulfate were used to match the N contents and provide sulfate application rates of 62 and 44 kg S[O.sup.2.sub.4]/ha for the DAP and SSP formulations, respectively. The total P K, S, Ca, Mg, and Na for the DAP and SSP fertiliser blends used to compare P exports were 8.8, 17, 5.1, 0.19, 0.22, and 0.42%, and 6.8, 12, 9.2, 17, 0.10, and 0.43%, respectively (Association of Official Analytical Chemists 1997). As is common in the area, a basal dressing of 71 kg K/ha was also applied. Blends were applied using a tractor and pre-calibrated, precision, drop spreader (Fiona D78, Scheby Maskinfabrik A/S, Bogense, Denmark).
Rainfall-induced overland flow study
The rainfall-induced overland flow study in 1998 used a pair of adjacent irrigation bays previously established for the purpose of monitoring nutrient exports on the Macalister Research Farm (38[degrees]0'S, 146[degrees]54'E) in the Macalister Irrigation District of south-eastern Australia. Pasture in the bays contained perennial ryegrass (Lolium perenne), white clover (Trifolium repens), and assorted invasive species including dock (Rumex spp.) and distichum (Paspalum paspaloides). The bays were c. 1.l ha (380 by 30 m) and had been laser-graded (1:400) 8 years prior to the experiment.
The soil type on both bays was a vertic, natric grey Sodosol (Isbell 1996), Natrixeralf (United States Department of Agriculture 1998), and Db2.43 (Northcote 1979). Soil was recovered to 100 mm by extracting 30 cores (6 x 5 across each of the 2 bays) and bulking. Following sampling, soils were air-dried and ground to pass a 2-mm sieve. Selected properties of the soils are presented in Table 1. Both bays were fitted with earthen ridges similar to check banks that directed overland flow through a 150-mm RBC Flume (Clemmens et al. 1984).
The bays were grazed 2 days prior to fertiliser application. DAP and SSP blends were applied to the respective bays 3 days after irrigation (c. 10 days prior to the next scheduled irrigation). However, a rainstorm of 18.3 mm resulted in overland flow occurring from both bays c. 5 h after fertiliser application. Overland flow was monitored using ISCO model 4230 bubbler flow meters and was automatically sampled using ISCO 3700 samplers (ISCO Inc., Lincoln, NE, USA). Rainfall was measured using an ISCO Model 674 rain gauge. Twenty-one and 18 samples per event were collected and analysed from the DAP and SSP treatments, respectively, at c. 5000 L (c. 0.5 mm) intervals for the single, rainfall-induced overland flow event that followed fertiliser application.
Irrigation-induced overland flow study
This study used 18 bays on 2 irrigation farms also in the Macalister Irrigation District; 8 bays were located on the Anderson farm (38[degrees]2'S, 146[degrees]47'E) and 10 bays on the Boulton farm (38[degrees]0'S, 146[degrees]53'E).
Eight months prior to the experiment, the 1.3-ha bays (c. 385 by 35 m) on the Anderson farm were laser-graded (1:400) and sown to a mixture of tall fescue (Festuca arundinacea cv. Advance), white clover (Trifolium repens cvv: Sustain and Will), and perennial ryegrass (Lolium perenne cvv. Aries and Yatsyn). The 1.2-ha bays (c. 340 by 34 m) on the Boulton farm were laser-graded (1:400) 6 months prior to the experiment and sown to a mixture of ryegrass (Lolium perenne cvv. Impact long rotation ryegrass and Dobson) and white clover (Trifolium repens cvv. Aran and Sustain). Bays that had been recently laser-graded (<12 months) were used to minimise differences in flow characteristics and background fertility between bays.
The soil on the Anderson farm was a eutrophic mottled-subnatric grey Sodosol (Isbell 1996), Natrixeralf (United States Department of Agriculture 1998), and Db2.43 (Northcote 1979). Soil on the Boulton farm was a melanic-sodic, eutrophic, grey Dermosol (Isbell 1996), Epiaquoll (United States Department of Agriculture 1998), and Gn3.92 (Northcotc 1979). Soil was recovered on each bay to 100 mm by extracting 30 cores (6 x 5) and bulking. Following sampling, soils were air-dried and ground to pass a 2-mm sieve. Analyses of the samples showed variation between bays on each farm was minimah Therefore the soil chemical properties are presented as means for each farm (Table 1).
In the absence of established monitoring equipment, on-bay sampling was used to compare nutrient exports. Sampling was initiated when the irrigation-wetting front passed strategic points located 50, 150, 250, and 350 m down the bays from the inlet. Overland flow was sampled by taking columns of water using the equipment described in Fig. 1 at the wetting front (i.e. within 2 m of the actual water front) where the depth of water varied between 30 and 70 mm. Samples were also taken at 100-m intervals behind the wetting front where the depth of water varied between 70 and 110 mm. For any sampling location, at least 14 columns of water were recovered, each e. 70 mL, and bulked to give one composite sample at each sampling point and time. In a number of instances, when the wetting front had reached 350 m, there was insufficient water at the 50 m mark to collect a sample.
The study commenced with irrigation of the Anderson farm on 21 October and 8 and 21 November 1999 and continued with irrigation of the Boulton farm on 22 November and 6 and 20 December of the same year. The elinaatic conditions during the study were typically variable, with a total of 7 mm of rainfall on 6 and 7 November and 3 mm on 21 November.
Water samples were collected over 3 consecutive irrigations on each farm, one before and two following fertiliser application. The treatment structure comprised 2 fertiliser treatments, DAP and SSP, and 4 fertiliser application times: 2, 6, and 10 days prior to the next irrigation for the Anderson farm and 1, 2, 6, and 10 days prior to irrigation for the Boulton farm. Each fertiliser timing treatment was applied to 2 irrigation bays on each farm, except for the 6 days treatment that was applied to 4 bays on each farm. Fertiliser and timing treatments were fully randomised to bays within farms.
Water samples were analysed for both TP and total dissolved P (TDP). Within 12 h of sampling, TDP samples were filtered using 0.45-[micro]m Millipore filters (Cat. No. HVLP04700, Millipore Corp., Bedford, MA, USA) and, along with the TP samples, stored at 4[degrees]C prior to analysis. Samples that could not be analysed within 5 days of sampling (<10%) were frozen at -20[degrees]C until processed. Samples were analysed using an alkaline persulfate digestion procedure adapted from Lachat Method number 30-115-01-1-B (Lachat Instruments 1997) on a Lachat QuikChem[R] 8000 (Lachat Instruments, Milwaukee, WI, USA) flow injection system using the traditional phosphomolybdenum blue chemistry for P determination. As TDP samples were prepared by filtration, the term dissolved refers only to materials that passed through a 0.45-[micro]m filter and does not imply that all materials in the filtrate were necessarily in solution. Certified reference material for P (Analytical Products Group Inc., Balpre, OH, USA), blanks, duplicate samples, and check standards were processed with the water samples and used as part of a Total Quality Control system to ensure the integrity of the data (Standards Australia 1999).
No statistical analyses of the rainfall-induced overland flow study were undertaken due to the lack of replication. In the irrigation-induced overland flow study, the TDP data were transformed [In(TDP + 1)] to comply with the assumption of equal variance of residuals across the dataset and to normalise the distribution of residuals. The transformation included the addition of a constant (1) to ensure all concentrations, including 0, were defined. A 2-level indicator was defined to distinguish between the first irrigation (for which no effect of fertiliser type or timing treatments was possible) and irrigations 2 and 3. Factorial effects of fertiliser type and timing by irrigation were then specified as being nested within the 'irrigation 2 and 3' level of this factor. This construct permitted data from the first irrigation, essentially a uniformity trial on bays, to enter into the analysis, augmenting the otherwise scant information available on error variance between bays. A mixed-effects model was defined using REML (Residual Maximum Likelihood) in GenStat (2000) including as fixed effects the fertiliser type and timing and irrigation treatment effects. nested within irrigations 2 and 3, in factorial combination with farm and a combination of 2 factors specifying the sample position in terms of the position of the water front and position behind the front. The random effects in the mixed-effects model were initially specified as sample nested within irrigation nested within bay. This was later simplified to sample nested within irrigation.bay, since the variance component estimate associated with between-bay variability was not significantly different from 0.
This simplification to the random effects model, in tandem with the exclusion of data corresponding to fertiliser timing of 1 day prior to irrigation (i.e. fertiliser applied 1 day before the second irrigation), rendered the design sufficiently balanced for analysis by ANOVA in GenStat (2000). To examine effects of fertiliser timing of I day prior to irrigation, a similar ANOVA was performed including data on all fertiliser timings, but restricting data to Boulton's in order to achieve the required design balance.
All data are presented as geometric means of the relevant treatments.
Results and discussion
Rainfall-induced overland flow study
The results from the rainfall-induced overland flow study are presented in Table 2. As in other studies of a similar scale (Bush and Austin 2001), despite thorough irrigation of the bays 3 days earlier and similar estimated steady-state infiltration rates, the volume of water exported in overland flow varied substantially between treatments making load comparisons potentially misleading. While the DAP-treated bay had a P concentration c. 40% higher than the bay to which SSP had been applied, the disparity in loads was 64% due to the greater flow from the DAP-treated bay. An inverse relationship between overland flow volume and nutrient concentration (i.e. a dilution effect) was not apparent. This suggests that under the conditions of this study, P mobilisation was limited by the rate at which P could be supplied to overland flow rather than quantity of P that was available for mobilisation (Nash et al. 2002).
Up to 53% of the 18.3 mm of rainfall that occurred c. 5 h after fertiliser application was collected as overland flow. The estimates of infiltration derived from water balance calculations are consistent with those measured on similar soils and the irrigation results from these bays (Nash et al. 2003b).
Irrigation-induced overland flow study
The treatment effects from the irrigation-induced overland flow study are summarised in Table 3. As TDP was >95% of TP and it is probable that most of the TDP is derived from fertiliser, only TDP data are presented.
Sampling location in relation to the position of the wetting front explained 49.7% of the total sum of squares (P < 0.001). TDP concentrations were highest in the wetting front and diminished exponentially with distance back up the bay (Fig. 2). As the wetting front represents that portion of overland flow that reaches the field edge, this is an important observation. Clearly, sampling irrigation water behind the wetting front or ponding water prior to sampling, as occurs with many flumes and weirs (Grant 1992) and occurred in the rainfall-induced overland flow study, may result in the underestimation of P exports.
There were significant differences between the P concentrations pre- and post-fertiliser application and between the first and second irrigations after fertiliser application (mean concentrations across all samples: 0.8, 4.6, and 1.6 mg TDP/L for irrigations 1, 2, and 3, respectively). In the wetting front at the lower end of the bay (i.e. 350 m), concentrations were of similar proportions and close to 4 times higher than mean concentrations (2.3, 17.6, and 6.5 mg TDP/L for irrigations 1, 2, and 3, respectively, Fig. 3).
Increased P export following fertiliser application has been observed in a number of studies (Olness et al. 1980; Nash et al. 2000; Withers et al. 2001), as has the decay in P export in the irrigations following fertiliser application (Austin 1998; Bush and Austin 2001). The decay in TDP concentrations following fertiliser application may reflect both sorption processes and the transportation of fertiliser P into the soil with irrigation water below the 'effective depth of interaction' (the depth of soil that overland flow interacts with) (Sharpley 1985; Ahuja 1986). In this study climatic conditions were conducive to dissolution of fertiliser granules because the dew point was exceeded on most evenings. It was therefore expected that sorption at the soil surface would reduce TDP concentrations in subsequent overland flow and the timing of fertiliser application prior to irrigation would explain a significant percentage of the total sum of squares. This was not the case. The fertiliser timing, represented by both the days between fertiliser application and overland flow (Irrl.Timing) and that factor's interaction with irrigation number post fertiliser application (Irrigation.Irrl.Timing) explained <0.9% of the sum of squares. The irrigation number following fertiliser application (Irrigation.Irrl) explained 15.2% of the sum of squares, suggesting that the physical transport of P into the soil may have been the more important of the 2 processes.
The effect of time between fertiliser application and irrigation was significant but variable, between farms and fertilisers. In keeping with other studies (Nash et al. 2000; Bush and Austin 2001), TDP concentrations tended to decrease as the interval between fertiliser application and irrigation increased (Fig. 4), but in this study the trend was inconsistent. The small percentage sum of squares explained by fertiliser timing probably reflects the importance of other effects such as irrigation number pre and post fertiliser application, sampling position behind the wetting front, and irrigation number post fertiliser application.
The effect of fertiliser compound on TDP in overland flow was highly significant (P < 0.001), but the difference between the 2 fertilisers was the reverse of the rainfall-induced overland flow study. The DAP treatments had lower TDP concentrations in overland flow than the SSP treatments (2.4 and 3.2 mg TDP/L, respectively, averaged across irrigations 2 and 3). Importantly, the TDP in the wetting front at 350 m showed a similar trend (8.7 and 13.4 mg P/L for the respective treatments). The rates at which P is mobilised from DAP and SSP granules, their reaction products, and the hydrology of the soils used in the rainfall-induced and irrigation-induced overland flow studies provide a possible explanation of these results.
In a laboratory-based incubation study using 2 soil types, P was mobilised more quickly from DAP than SSP particularly at soil moisture contents >15% (Nash et al. 2003a). For example, after 8 h at c. 20% soil moisture, 17.6% of the P [In(%P) 2.87, s.e. 0.079] was extracted from DAP applied to soil, while under similar conditions only 5.3% of the P [ln(%P) = 1.67, s.e. 0.079] was extracted from the SSP. In the same study >50% of the P in DAP was released within 1 h of immersion in slowly oscillating water, while 50% release took 6 h for SSP.
In the rainfall-induced overland flow study there was c. 5 h between fertiliser application and overland flow and only c. 10 mm of the rainfall infiltrated the soil. With limited time for soil and fertiliser interaction before flow began, the P mobilisation probably reflected the results of the immersion experiments (i.e. DAP > SSP) (Nash et al. 2003a). With the soil being wet from the previous irrigation, the infiltration rates would be more uniform than if the soil had been drier (i.e. nearer wilting point) prior to the rain (Austin 1998). In addition, the concentration of P entering soil would approximate that of the overland flow, as P would be well mixed through the water column by raindrop impact on the recently grazed pasture (Ahuja and Lehman 1983). Consequently, while the P concentrations in the rainfall-induced overland flow study may have been slightly reduced by infiltration of additional P from DAE P exports in drainage probably reflected differences in the short-term water-extractable P between the fertilisers (Nash et al. 2003a).
In the irrigation-induced overland flow study, P exports are also likely to have reflected P mobilisation. In border irrigation systems, water is applied when the soil water deficit (i.e. evaporation--rainfall; measured using Class A pan evaporation) is c. 40 mm. As a consequence, the infiltration rate decays rapidly behind the wetting front (Austin 1998). Under these circumstances, P rapidly mobilised from DAP would infiltrate the soil as the wetting front passed. While some P from SSP would also infiltrate the soil, much of the P from SSP would be mobilised when infiltration rates had declined to near steady-state values, behind the wetting front.
The concentrations of P behind the wetting front support this explanation. In the event that P from DAP was rapidly mobilised from the granules, the P concentrations behind the wetting front should be low and close to uniform, as P from the granule and dissolution products would be depleted. In the case where P mobilisation occurs more slowly, as has been postulated for SSP the concentration of P should be inversely related to the distance behind the wetting front, commensurate with granule to water contact time. Figure 2 suggests that such a relationship may exist in these data when samples were taken 350 m from the inlet.
Interestingly, the P concentration in the wetting front of the irrigation-induced overland flow was higher for the SSP than the DAP treatment. The fertilisers were generally applied well before irrigation (i.e. >24 h). Laboratory studies (Nash et al. 2003a) suggest that sufficient time had elapsed for a significant portion of the P to be mobilised in a water-extractable form from the DAP, and to a lesser extent the SSP, granules. This P would infiltrate the soil as the true wetting front passed. By necessity, overland flow was sampled where the water depth varied from 30 to 70 mm (i.e. within 2 m of the actual water front). Consequently, the estimates of P are unlikely to reflect the concentrations infiltrating at the true wetting front.
Environmental and management implications
It is tempting when presented with a set of laboratory data that suggest considerably more P is mobilised from one P fertiliser compared with another to apply a blanket recommendation to include in 'Best Management Practice' guidelines (Waters 1996). These experiments demonstrate the complexity of real farming systems and the complex relationships between physical and chemical phenomenon that govern their off-site impact.
It would appear that if fertiliser is to be applied, substantial reductions in dissolved P exports (c. 25%) may be possible by avoiding the application of DAP to recently wet soil. in rainfed systems without irrigation this could be achieved by avoiding the use of DAP in winter and spring. However, the evidence is, as yet, insufficient to support the use of a legislative instrument to achieve the desired outcome. Firstly, it would be premature to base such a recommendation on laboratory data and limited field study. Secondly, for reasons of trafficability and pasture damage, few fertilisers are applied to truly saturated soil. Depending on the soil moisture and infiltration rates, avoiding the use of DAP may well not reduce P export compared with SSP. Where a composite fertiliser is required, an additional problem with prohibiting the application of DAP to moist soils is the long-term damage to soil structure caused by additional passes of the spreading equipment. If this results in reduced infiltration rates, then it may increase infiltration excess overland flow and nutrient exports.
The situation for irrigation farms is marginally clearer. Phosphorus concentrations in the wetting front at 350 m (14 and 22 mg P/L for DAP and SSP, respectively, for the irrigation following fertiliser application) suggest that the exclusive use of DAP may reduce P exports compared with SSP. The wetting front is primarily responsible for P exports and was sampled c. 2 m behind the leading edge. Assuming that the measurements made in this zone reflect the concentration profiles from the two fertilisers adequately, it is reasonable to assume that in the absence of rainfall the use of DAP will reduce P exports. However, it would be prudent to test this assumption under a range of conditions before it is universally applied. Furthermore, the mechanisms that lead to differences in P concentrations between these two products should be established to support conclusions drawn from these empirical studies.
The time between fertiliser application and overland flow appears to be a key management practice that can be used to reduce P exports. In a rainfed situation, fertilisers can be applied in late summer/early autumn when overland flow is unlikely. In addition, fertiliser application schedules can be aligned with weather forecasts to minimise environmental impacts in late winter/early spring. These experiments suggest that on border irrigation farms, maximising the time between fertiliser application and subsequent irrigation may reduce P exports by c. 30-50% compared with irrigation immediately following fertiliser application. However the decrease in P exports is likely to be variable and probably will depend on soil type and conditions. As there do not appear to be any adverse agronomic effects of this practice, maximising the time between fertiliser application and irrigation could reasonably be considered a part of prudent farm management.
Table 1. Selected surface characteristics (0-100 mm) of the soils on the farms used for comparing P exports from single superphosphate and diammonium phosphate Characteristic Macalister Anderson Boulton Research Farm Farm Farm Textural class Clay loam Fine sandy Light clay (very fine clay loam (very fine sandy) sandy) pH(water) (A) 5.4 5.7 5.6 pH(Ca[Cl.sub.2]) (B) 4.8 5.2 4.7 Exch. Al ([cmol.sub.c]/ 14 <10 -10 kg) (C) EC (dS/m) 0.19 0.28 0.20 (1:5 soil:water) (D) Olsen P (mg/kg) (F) 31 32 14 Skene K (mg/kg) (F) 224 100 125 Available S (mg/kg) (G) 410 35 27 Organic matter (%) (H) 5.7 6.2 8.6 Extractable Fe (%w/w) (I) 0.35 0.24 0.83 Extractable Al (%w/w) (I) 0.08 0.044 0.14 P Adsorption (mg/kg) (J) 11 16 47 Exch. cations 6.9 9.3 18.1 ([cmol.sub.c]/kg) (K) Ca 4.8 7.7 12.4 Mg 1.6 1.3 5.2 Na 0.1 <0.1 0.3 K 0.4 0.2 0.3 Rayment and Higginson (1992): (A) 4A1, (B) 4B2, (C) 15G1 (1 M KCl extract), (D) 3A1, (E) 9C2, (F) 18B1, (G) Peverill (1976). Rayment and Higginson (1992): (H) 6A1, (I) 13A1, (J) 9J1 at 5 mg P/L equilib. conc., (K) 15D3 (ammonium acetate). Table 2. Selected storm, site and overland flow characteristics from the rainfall-induced overland flow study of fertilisers Treatment: DAP (A) SSP (B) Chemical characteristics Total P (mg/L) 97 70 Total dissolved P (C) (mg/L) 89 63 P export (kg/ha) 9.0 5.5 Physical characteristics Incident rainfall (mm) 18.3 18.3 Duration of overland flow (h) 38 38 Actual overland flow (mm) 9.3 7.9 Estimated infiltration (mm) 9.0 10.4 Overland flow (%) 53 43 Estimated steady-state infiltration rate (mm/h) (D) 0.24 0.27 (A) Diammonium phosphate. (B) SuPerfect[TM] (Incitec Pivot Ltd Geelong, Australia). (C) Defined as P < 0.45 [micro]m by filtration. (D) Estimated as (rainfall - overland flow/overland flow duration). Table 3. Analysis of variance for ln(TDP+1) data from the irrigation-induced overland flow study, excluding fertiliser timing of 1 day prior to irrigation 2 The factor 'Irr1' distinguishes between the first irrigation (for which no effect of fertiliser type or timing treatments was possible, as fertiliser was only applied between irrigations 1 and 2) and irrigations 2 and 3. Factorial effects of fertiliser type and timing by irrigation were nested within the 'irrigation 2 and 3' level of 'Irr1'. Thus, for example, 'Irr1.Fert', measures the main-effect of fertiliser type, over irrigations 2 and 3 Sources of variation (A) d.f. (B) Total SS (C) MS (D) P (E) Farm.paddock.irrigation stratum Farm 1 0.31 1.10 0.002 Irr1 1 20.48 73.21 <0.001 Farm.Irr1 1 0.73 2.61 <0.001 Irrigation.Irr1 1 15.22 54.42 <0.001 Irr1.Timing 2 0.31 0.55 0.008 Irr1.Fert 1 1.06 3.80 <0.001 Farm.Irrigation.Irr1 1 0.89 3.20 <0.001 Farm.Irr1.Timing 2 0.73 1.31 <0.001 Irrigation.Irr1.Timing 2 0.58 1.04 <0.001 Farm.Irr1.Fert 1 0.04 0.16 0.205 Irrigation.Irr1.Fert 1 0.08 0.30 0.086 Irr1.Timing.Fert 2 0.09 0.15 0.21 Farm.Irrigation.Irr1.Timing 2 0.26 0.46 0.015 Farm.Irrigation.Irr1.Fert 1 0.05 0.17 0.183 Farm.Irr1.Timing.Fert 2 0.27 0.48 0.013 Irrigation.Irr1.Timing.Fert 2 0.01 0.02 0.836 Five factor interactions 2 0.04 0.07 0.483 pooled Residual 24 0.62 0.09 Farm.paddock.irrigation.sample stratum Front.Position 10 49.73 17.77 <0.001 Farm.Front.Position 9 0.37 0.15 <0.001 Irr1.Front.Position 10 4.98 1.78 <0.001 Farm.Irr1.Front.Position 9 0.48 0.19 <0.001 Irrigation.Irr1.Front. 9 3.47 1.38 <0.001 Position Irr1.Timing.Front.Position 18 0.40 0.08 <0.001 Irr1.Fert.Front.Position 10 0.49 0.17 <0.001 Five factor interactions 81 1.46 0.07 <0.001 pooled Residual 293 2.07 0.03 (A) Variate: ln(TDP + 1). (B) Degrees of freedom. (C) Percentage of total sum of squares. (D) Mean sum of squares. (E) P-value.
The authors thank the State Government of Victoria, GippsDairy, Dairy Australia Ltd, and Incitec Pivot Limited for their financial support of this research. We also thank the Soils and Water Team from PIRVic Ellinbank and the Macalister Research Farm Co-operative Ltd, Graeme and Chris Anderson, and Ken, Alison and Peter Boulton on whose farms this work was undertaken.
Ahuja LR (1986) Characterization and modeling of chemical transfer to runoff. Advances in Soil Science 4, 149-188.
Ahuja LR, Lehman OR (1983) The extent and nature of rainfall soil interaction in the release of soluble chemicals to runoff. Journal of Environmental Quality 12, 34-40.
Association of Official Analytical Chemists (1997) 'Official methods of analysis of AOAC International.' (AOAC International: Gaithersberg, MA)
Austin NR (1998) Sustainable use of water and fertiliser in the irrigated dairy industry of Australia. PhD thesis, University of Melbourne.
Bush BJ, Austin NR (2001) Timing of phosphorus fertilizer application within an irrigation cycle for perennial pasture. Journal of Environmental Quality 30, 939-946.
Clemmens AJ, Bos MG, Replogle JA (1984) Portable RBC flumes for furrows and earthen channels. Transactions of the American Society, of Agricultural Engineers 27, 1016-1026.
Connor DJ, Smith DF (Eds) (1987) 'Agriculture in Victoria.' (Australian Institute of Agricultural Science: Parkville, Vic.)
Edwards DR, Daniel TC (1994) Quality of runoff from fescuegrass plots treated with poultry litter and inorganic fertilizer. Journal of Environmental Quality 23, 579-584.
Environment Protection Authority (1995) Protecting water quality in Central Gippsland. Schedule F5--Waters of the Latrobe and Thomson River Basins and Merriman Creek Catchment and Draft Policy Impact Assessment. Environment Protection Authority, 444, Melbourne, Vic.
GenStat (2000) 'GenStat for Windows.' Release 4.2. 5th edn (VSN International Ltd: Oxford)
Grant DM (1992) 'ISCO open channel flow measurement handbook.' (ISCO Inc.: Lincoln, NE)
Gutteridge, Haskins & Davey (1992) An investigation of pollution in the Murray-Darling River system. Report for Murray-Darling Basin Commission, Canberra.
Isbell RF (1996) 'Australian Soil Classification.' (CSIRO Publishing: Melbourne, Vic.)
Lachat Instruments (1997) 'Methods list for QuikChem automated analyzers.' (Lachat Instruments: Milwaukee, WI)
Nash DM, Halliwell D, Cox J (2002) Hydrological mobilisation of pollutants at the slope/field scale. In 'Agriculture, Hydrology and Water Quality'. (Eds PM Haygarth, SC Jarvis) pp. 225-242. (CABI Publishing: Oxon, UK)
Nash DM, Hannah M, Clemow L, Halliwell D, Chapman D (2003a) A laboratory study of phosphorus mobilisation from commercial fertilisers. Australian Journal of Soil Research 41, 1201-1212.
Nash DM, Hannah M, Halliwell D, Murdoch C (2000) Factors affecting phosphorus export from a pasture based grazing system. Journal of Environmental Quality 29, 1160-1165.
Nash DM, Savage G, Clemow LM, Webb BJ (2003b) Macalister Research Farm water balance. Department of Primary Industries, Ellinbank, Vic.
Nexhip KJ, Mundy GN, Collins MD, Austin NR (1997) Development of nutrient water quality targets for irrigated pasture sub-catchments. Institute of Sustainable Irrigated Agriculture, 15037, Tatura, Vic.
Northcote KH (1979) 'A factual key for the recognition of Australian soils.' (Rellim Technical Publications: Glenside, S. Aust.) Olness AE, Rhoades ED, Smith SJ. Menzel RG (1980) Fertilizer nutrient losses from rangeland watersheds in central Oklahoma. Journal of Environmental Quality 9, 81-86.
Peverill KI (1976) The availability of phosphorus and sulphur in soil. PhD thesis, University of Melbourne.
Preedy N, McTiernan K, Matthews R, Heathwaite L, Haygarth P (2001) Rapid incidental phosphorus transfers from grassland. Journal of Environmental Quality 30, 2105-2112.
Rayment GE, Higginson FR (1992) 'Australian laboratory handbook of soil and water chemical methods.' (Inkata Press: Melbourne, Vic.)
Sharpley AN (1985) Depth of surface soil-runoff interaction as affected by rainfall, soil slope, and management. Soil Science Society of America Journal 49, 1010-1015.
Sharpley AN (1997) Rainfall frequency and nitrogen and phosphorus runoff from soil amended with poultry litter. Journal of Environmental Quality 26, 1127-1132.
Sharpley AN, Syers JK, Gregg PEH (1978) Transport in surface runoff of phosphorus derived from dicalcium phosphate and superphosphate. New Zealand Journal of Science 21, 307-310.
Standards Australia (1999) 'Australian Standard General requirements for the competence of testing and calibration laboratories.' Standards Australia International Ltd, AS ISO/IEC 17025-1999, Strathfield, NSW.
United States Department of Agriculture (1998) 'Keys to Soil Taxonomy.' (United States Department of Agriculture Natural Resources Conservation Service: Washington, DC)
Waters CT (1996) 'Dairy land manager.' (Department of Natural Resources and Environment, Melbourne, Vic.)
Webster IT, Parslow JS, Grayson R, Molloy RP, Andrewartha J, Sakov E Tan KS. Walker SJ, Wallace BB (2001) Gippsland Lakes enviromnental study--Assessing options for improving water quality and ecological function. CSIRO Land and Water, Glen Osmond, S. Aust.
Withers PJA, Clay SD, Breeze VG (2001) Phosphorus transfer in runoff following application of fertiliser manure and sewage sludge. Journal of Environmental Quality 30, 180-188.
D. Nash (A,C), M. Hannah (A), L. Clemow (A), D. Halliwell (A), B. Webb (A), and D. Chapman (B)
(A) primary Industries Research Victoria (PIRVic), RMB 2460 Hazeldean Rd, Ellinbank, Vic. 382 l, Australia.
(B) Institute of Land and Food Resources, University of Melbourne, Parkville, Vic. 3010, Australia.
(C) Corresponding author; email: David.Nash@dpi.vic.gov.au
Manuscript received 12 May 2003, accepted 1 December 2003
|Printer friendly Cite/link Email Feedback|
|Author:||Nash, D.; Hannah, M.; Clemow, L.; Halliwell, D.; Webb, B.; Chapman, D.|
|Publication:||Australian Journal of Soil Research|
|Date:||May 1, 2004|
|Previous Article:||Comparison of extraction techniques for measuring exchangeable cations in calcareous soils.|
|Next Article:||The liming effect of bauxite processing residue (red mud) on sandy soils.|